Seakeeping for Autonomous Ship in Heavy Weather

I recognize this is a bit of an old thread, and I welcome any redirect to a more current conversation.

I recently started working for the Navy on AI for autonomous navigation of mid-sized to larger vessels (100’ to 1000’) in threateningly high seas (8m or greater wave height), and I was wondering, based on the experience of mariners who have been through it, what would be the most important information to have about the environment and the ship?

Clearly, having statistics about the mean swell direction and height is not enough, you need to know if the ship has the power to maintain steering-way in that sea and what is the best course to cut through the wave field. But I also wonder how actively the ship is being steered or throttled, and what other information from the environment is being used. For example, is the “real local” swell direction moving around and the ship has to constantly adjust course to cut the best path into the waves under power?

I am an AI guy and an oceanographer secondarily, but not a ship’s captain by any stretch, so any ideas are helpful!

1 Like

In the worst weather the ship is taken off of autopilot and a heading is maintained in the direction the skipper or OOTW thinks puts the least amount of strain on the vessel. In most cases this is simply a matter of looking out the windows and seeing which way the primary sea is running, and gauging how badly the ship is moving under your feet, then trial and error to adjust course, until the ship rides easier. This could go on for days, with frequent minor course changes, and on smaller vessels, constant adjustment of throttle.

An important thing to remember is this: what the ship can take and what the cargo can take are two different things. A ship can easily survive a storm but come back with damaged cargo because of heavy rolling, etc. But it’s not just the degree of motion or period of the roll. There is an acceleration aspect to the motion which can play havoc with cargo, and play a key role in determining when a captain needs to change course.

Stability, the rolling period of ship and period of the waves is important also stress, bending moments.

Likely the ship will have sufficient power. The issue is damage or extreme motions that might result from using excessive power. Sometime becomes an issue of having to slow to the point where the ship lose steerage.

Wind speed and direction may also be a factor in what course and speed can be maintained.

Tell him nothing; he is planning to take our jobs!


This is spot on. You can have a heading and speed set that is working fairly well and then a front passes and the wind shifts, or the predominant swell is overtaken by another from a different bearing and an adjustment is necessary. There is a lot of it that is just how you feel in the seat of your pants.

Thanks to everyone for your input.

That’s interesting. I wonder if anyone has come up with a measurement system to try to quantify “how badly the ship is moving under your feet”, by which I assume you mean flex and give in the mechanical superstructure of the vessel? There is also the idea of the ship not going the direction it is being steered due to the force of the wind and the waves, but I guess that is more of a heading thing than “under foot”.

Also, you talk about throttle adjustments, but to what end? Is it a matter of trying to keep constant speed and power going up and down these large hills of water?

Luckily, cargo is not a big part of what I am doing, unlike in the merchant navy, though I am sure if they have ammo on board, they will want it tied down very well!

Also I sympathize with the notion that automated ships could loose people jobs. My particular work is all about navigation in high seas for ship safety. As someone who has spent a lot of my career doing machine monitoring and maintenance work, I just can’t see long-run unmanned cargo ships making much sense financially. The ship has to run to make money, and running is when things need maintenance and break down, so that is when tending and fixing them has to happen. So, until we have machines that diagnose and fix themselves, which is a very long way off, I feel like unmanned merchant fleets don’t make much sense. The military, of course, has much different missions, and the main goal of autonomy is keeping men out of harm’s way, like in mined waters. Plus for the Navy, two days in overhaul for every day out is common, but would never be financially viable in commercial shipping.

Thanks for your help!

1 Like

They have: “ Although considerable efforts can be made in predictions of ship wave loads during prototype design stage, there is still a real need for real-time monitoring of the stress state of ships during their service life. A set of structural safety monitoring system, which is intended to be used for full-scale ships’ hull stress monitoring and evaluation, has recently been developed by our research team. The stress monitoring system was fitted onboard the large-scale model for a preliminary trial and validation.”

The paper covers a lot of relevant ground to your interests.


The Theory and Practice of Seamanship

4 posts were split to a new topic: Storm Oil

What, this must be a joke…

In my experience on a small coast-wise ship while hove-to or nearly hove-to in head seas, as the ship climbs the face of a large wave, when the crest of the wave is reached, or just before, the throttles are pulled back to give the ship more time to pivot over the crest of the wave so as not to slam on the back side of the wave. Once over the crest the throttle can be increased again as needed to maintain steerage to meet the next wave.

On larger, heavier ships, generally speaking the response is more delayed so while rpm might be adjusted to keep the speed within range (between loss of steerage and heavy slamming/pounding) but typically course and speed adjustments would be made less frequently.

This is not a joke. 2 to 1 is something an aircraft maintenance guy told me as a rule of thumb. If you look at something like this:

Anywhere it says “home port”, “sea trials”, or “selected restricted availability (SRA)” that means “not deployed”. You will see it is most of the time. Generally beyond 2 to 1 if there is no naval aggression going on.

Now, I don’t want to make the Navy sound bad, if bad things are happening, they can stay active way beyond 2 to 1, I just brought it up to emphasize that applications in the Navy in autonomy does NOT mean much regarding autonomous merchant marine operations.

I see what you mean, quite a difference with the merchant marine where there is sometimes hardly time to spare to carry out necessary maintenance, cleaning/painting of the hull and cleaning of the propellers etc.

BTW, you could search gCaptain with the term ‘unmanned ships’. You will find a number of topics about this subject popping up which could contain useful information for you.

Merchant ships are losing money when not moving. With the military at-sea operations are going to cost more than in-port.

One of ours, not sure if it is the Dauntless or the Daring, has never actually sailed at all. Ever.

I believe two man hours maintenance to one flight hour would be considered a utopian figure for a fighter plane. See below (from user HaveBlue on Fighter Maint. ManHours Per FlightHour Comparison -

But it’s a lot more complicated for a ship because people live on the thing, and counting man hours maintenance per hour at sea isn’t, IMO, fairly represented by non-deployed time. I imagine it’s a number the Navy could cough up for a given class of ship if it wanted to, once the definitions were settled as to what counts as maintenance. And of course it would be huge, because even little ships are big and Navy ships are jam-packed with fancy gear.

If you’re just counting calendar days, I don’t think it’s fair to count home-port or alongside time as maintenance just because they’re not at sea.

Got these from the Air & Space magazine January 2008:

Early F-117…- 113 to 1
Concorde…- 18 to 1

Here’s the compilation from this thread so far:

Saab Draken.- 50 to 1
Eurofighter…- 9 to 1
F-14… - 24 to 1
F-18E/F…- 6 to 1
F-18E/F…- 15 to 1 (different source)
Saab Gripen…- 10 to 1

C-17…- 20 to 1
F-15A/B…- 32.3 here thru f117 stats from (HaveBlue and the F-117A by David Aronstein)
F-15C/D…- 22.1
F-16A…- 19.2
F-117…- 150 (pre 1989)
F-117…- 45 (after improvements, post 1989)
CH-46E…- 19.6 in 1995
CH-46E…- 27.2 in 2000
CH-53D…- 24.8 in 1995
CH-53D…- 27.9 in 2000
F-20…- 5.6 (
A-6E…- 51.9 DMMH/FH (FA-18 vs A-6 (Bill DuBois))
F/A-18C…- 19.1 DMMH/FH
B-2…- 124

“The actual B-2 maintenance man-hours per flying hour at Whiteman Air Force Base averaged 124 hours over 12 months ending in March 1997.”

Mirage 2000…- 10 Dash 5 (Maintenance)
Gripen…- 12 (Maintenance)

Interesting about the A-7A from a 1964 article…

The contract
between the Navy and Ling-Temco-Vought calls for an 80-per-cent
probability that the aircraft will all achieve mission success, and that
maintenance man-hours per flight hour must not exceed 11.5 or a
penalty will be imposed. If maintenance man-hours per flight hour
reach 13, the contractor must pay the Navy a penalty of $50 per
hour; if the figure reaches 17 the Navy is to receive $700 per hour.
If the maintenance requirement is higher still, then the airplane
will be returned to the contractor for a complete refund of its cost
to the Navy."

Thanks to everyone for your posts on my behalf. I am working through all the references and old posts on other threads to get myself up to speed, and then I am sure I will have more questions! I have “Theory and Practice for Seamanship” on order from Amazon, and I am sure that will bring up even more questions!

One more question that might be interesting to the group: anybody know of a novel that includes heavy weather trials and decision makings for large-ish vessels? Even age of sail books might be useful if they get into what choices the crew need to make. On the other hand, something like “A Perfect Storm” is more about the drama of the rescue and the weather event (and smaller ships). Wondering if there is a “storm waves classic” out there someone has read!

On 2 to 1, I agree it is oversimplified, but I think there is a good point there that the Navy and merchant missions and cost issues are very different. (also I wanted to correct that I said “aircraft maintenance” when I meant “aircraft carrier maintenance”. I was on the Ronald Reagan carrier between Guam and Hawaii and it was an amazing experience).

On oil calming waves, I don’t really want to jump into that! From the point of view of physics, it is hard to see any surface effect doing much to large waves that have traveled miles to reach a ship. And those are the ones I am interested in.

Thanks again to all for your input.

Into The Raging Sea by Rachel Slade.
Many discussions on here about it.

Thanks, this is very helpful!